Besides, basic packet forwarding from a start node to an end node of a network, data path functions have recently been provided. While being forwarded from node to node of the network, data is processed according to given service-related functions. This has become possible thanks to programmable data path processors, evolved specialized coprocessors etc. The number and type of such data path functions which are scattered over network nodes is growing rapidly with the deployment of new equipment.
In these new developments, forwarding and transforming data is related to a service. The service is provided by one or more functions which are applied to the data flowing through the network. In order to provide a useful service, the functions applied to the data must be executed in nodes of the network, wherein the data is transmitted through the nodes in a specifically optimized manner.
Increasingly, the service depends on specific customer needs and requires a certain number of functions, the combination of which is not known a priori. As performing these functions requires different capacities within different nodes, the need to optimize the distribution and usage of such functions within the network nodes is coming more and more into focus.
Especially the complexity introduced by the growing number of network nodes mainly in hierarchical networks and the increasing variety of functions to be provided to the forwarded data is complicating the optimization of the usage of the service within the network.